Supply Chain Risk Management

203

11

Using Big Data and Analytics

to Manage Risk

By now you’ve probably heard about or have had some experience with

something called “big data.” While we may have heard of the concept, tak-

ing advantage of the treasure trove of data that resides at most companies

remains an evolving challenge. With an estimate of more than 15 million

gigabytes of new information collected every day (15 petabytes), which is

eight times the information in all U.S. libraries, it’s no wonder most com-

panies are wondering how to use big data to their advantage.

1

But is using big data going to be that straightforward? A report titled

Big Data Insights and Innovations Report revealed some ndings that

relate directly to big data and its uses.

2

First, many organizations are chal-

lenged by data overload and an abundance of trivial information. And

important data are not reaching practitioners in ecient time frames.

Current technology is also not yet at the level of providing measurable,

reportable, and quantiable data in areas including production sched-

uling, inventory, and customer demand across the entire supply chain.

Furthermore, despite the sophistication of current systems, data are not

always easily accessible to internal users. Finally, noticeable gaps are

present in many end- to- end supply chain ow models. Other than these

“minor” issues, everything is working just ne in the world of big data and

risk management.

In this chapter we’ll advance some denitions and an overview of big

data and predictive analytics; talk about the process for successfully lever-

aging big data; present barriers and challenges with big data; and present

tools, techniques, and methodologies that support big data and analytics.

e chapter concludes with examples of companies using big data and how

these companies are leveraging their data to help manage supply chain risk.

204 • Supply Chain Risk Management: An Emerging Discipline

WHAT IS BIG DATA AND PREDICTIVE

ANALYTICS, REALLY?

To some observers big data got its start around 2003 with the advent of the

Data Center Program at Massachusetts Institute of Technology (MIT).

3

Before this, most of the early research in the late 1990s used the term data

analytics as a key descriptor. It becomes critical to dene the terms big

data and predictive analytics.

According to the Leadership Council of Information Advantage, big

data is not a precise term. is group sees it as data sets that are grow-

ing exponentially and that are too large, too raw, or too unstructured for

analysis using relational database techniques. So, where is this unbeliev-

able amount of unstructured data coming from? According to one source,

the amount of data available is doubling every two years and is emanat-

ing from not only traditional sources but also industrial equipment, auto-

mobiles, electrical meters, and shipping crates, just to name a few. e

information gathered includes parameters such as location, movement,

vibration, temperature, humidity, and chemical changes in the air.

4

Predictive analytics (PA) encompasses a variety of techniques from sta-

tistics, data mining, and game theory that analyze current and historical

facts to make predictions about the future. In business, predictive models

exploit patterns found in historical and transactional data to identify risks

and opportunities. Models capture relationships among factors to allow

assessment of risk or potential associated with a particular set of condi-

tions, guiding decision making for specic transactions.

5

Predictive analytics has been traditionally used in actuarial science,

nancial services, insurance, telecommunications, retail, travel, health

care, and pharmaceuticals. Yet it is barely mentioned in the manufacturing

and supply chain arenas. One of the best known and early applications of

PA is credit scoring, which is used throughout nancial services. Scoring

models process customers’ credit history, loan applications, customer data,

and so forth, in an eort to rank- order individuals by their likelihood of

making future credit payments on time. A well- known example is the

FICO score.

IBM, a leading provider of big data systems, maintains that more than

90% of the data that exists in the world today was created within the last

two years. We are in an age where more than 2.5 quintillion bytes of data